diff --git a/frontend/src/components/ContextLimitDialog.tsx b/frontend/src/components/ContextLimitDialog.tsx
index 3ed7b32ae..9ff36470c 100644
--- a/frontend/src/components/ContextLimitDialog.tsx
+++ b/frontend/src/components/ContextLimitDialog.tsx
@@ -62,8 +62,8 @@ export function ContextLimitDialog({
                 <li className="flex items-start gap-2">
                   <MessageCircle className="h-4 w-4 mt-0.5 shrink-0" />
                   <span>
-                    <strong>Switch to a model with more context</strong> - Try DeepSeek R1, Mistral,
-                    or other models that support 128k tokens
+                    <strong>Switch to a model with more context</strong> - Try DeepSeek R1 or other
+                    models that support 128k tokens
                   </span>
                 </li>
               )}
diff --git a/frontend/src/components/ModelSelector.tsx b/frontend/src/components/ModelSelector.tsx
index d542c919c..ce3d6c5f6 100644
--- a/frontend/src/components/ModelSelector.tsx
+++ b/frontend/src/components/ModelSelector.tsx
@@ -82,21 +82,6 @@ export const MODEL_CONFIG: Record<string, ModelCfg> = {
     requiresPro: true,
     tokenLimit: 128000
   },
-  "mistral-small-3-1-24b": {
-    displayName: "Mistral Small 3.1 24B",
-    shortName: "Mistral Small",
-    badges: ["Pro"],
-    requiresPro: true,
-    supportsVision: true,
-    tokenLimit: 128000
-  },
-  "qwen2-5-72b": {
-    displayName: "Qwen 2.5 72B",
-    shortName: "Qwen 2.5",
-    badges: ["Pro"],
-    requiresPro: true,
-    tokenLimit: 128000
-  },
   "qwen3-coder-480b": {
     displayName: "Qwen3 Coder 480B",
     shortName: "Qwen3 Coder",
diff --git a/frontend/src/components/UpgradePromptDialog.tsx b/frontend/src/components/UpgradePromptDialog.tsx
index bf3a8fadc..9fd896338 100644
--- a/frontend/src/components/UpgradePromptDialog.tsx
+++ b/frontend/src/components/UpgradePromptDialog.tsx
@@ -193,7 +193,7 @@ export function UpgradePromptDialog({
         benefits: [
           "All models run in secure, encrypted environments",
           "Access to DeepSeek R1 for advanced reasoning",
-          "OpenAI GPT-OSS, Mistral, Qwen, and more",
+          "OpenAI GPT-OSS, Qwen, and other advanced models",
           "Higher token limits for longer conversations",
           "Priority access to new models as they launch"
         ]